Paper submitted to APGCE 2022 Kuala Lumpur, Malaysia, 28 – 29 November, 2022
Abstract
In shallow water environments where usable offsets are limited, Full Waveform Inversion (FWI) has become a tool-of-choice to generate high-resolution velocity updates. Traditional approaches such as refraction and reflection traveltime tomography can struggle with capturing sharp lateral and vertical velocity variations responsible for structural distortions deeper in the section. Classical implementation of FWI works at its best when very low frequencies and long offsets are recorded. This allows the inversion scheme to start from a smooth initial velocity model. However, the absence of quality low frequencies below 4 Hz and offsets limited to 6000 m in this project meant that the initial model built from legacy stacking velocities had to be improved prior to running FWI.
We first demonstrate a methodology to automatically detect the top of the shallow carbonate platforms and pinnacle reefs present near the seabed. The small-scale velocity variations added to the starting model bring immediate structural improvement down to the reservoir level and enable FWI to start at a higher frequency band. Alternating between passes of FWI and conventional reflection tomography leads to an updated velocity model conforming to the complex geology. The resulting depth migration enables better fault positioning and mapping of key reservoir horizons.